March 9, 2007 10:44 WSPC/139-IJMPA 03619

1st Reading

1 International Journal of Modern Physics A

Vol. 22, No. 00 (2007) 1–24
3 c World Scientific Publishing Company

ON (2 + 2)-DIMENSIONAL SPACE TIMES,

5 STRINGS AND BLACK HOLES

C. CASTRO
7 Center for Theoretical Studies of Physical Systems,
Clark Atlanta University, Atlanta, GA 30314, USA
9 castro@ctsps.cau.edu

J. A. NIETO
11 Facultad de Ciencias Fı́sico-Matemáticas de la Universidad Autónoma de Sinaloa,
80010, Culiacán Sinaloa, México
13 nieto@uas.uasnet.mx

Received 9 November 2006

15 Revised 5 January 2007

We study black hole-like solutions (space–times with singularities) of Einstein field equa-
17 tions in 3 + 1 and 2 + 2 dimensions. We find three different cases associated with
hyperbolic homogeneous spaces. In particular, the hyperbolic version of Schwarzschild’s
19 solution contains a conical singularity at r = 0 resulting from pinching to zero size r = 0
the throat of the hyperboloid H2 and which is quite different from the static spherically
21 symmetric (3 + 1)-dimensional solution. Static circular symmetric solutions for metrics
in 2 + 2 are found that are singular at ρ = 0 and whose asymptotic ρ → ∞ limit leads to
23 a flat (1 + 2)-dimensional boundary of topology S 1 × R2 . Finally we discuss the (1 + 1)-
dimensional Bars–Witten stringy black hole solution and show how it can be embedded
25 into our (3 + 1)-dimensional solutions. Black holes in a (2 + 2)-dimensional “space–time”
from the perspective of complex gravity in 1 + 1 complex dimensions and their quater-
27 nionic and octonionic gravity extensions deserve furher investigation. An appendix is
included with the most general Schwarzschild-like solutions in D ≥ 4.

29 Keywords: Strings; black holes; 2 + 2 dimensions; general relativity.

PACS numbers: 04.60.-m, 04.65.+e, 11.15.-q, 11.30.Ly

31 1. Introduction
Through the years it has become evident that the 2 + 2 signature is not only
33 mathematically interesting1,2 (see also Refs. 3–5) but also physically. In fact, the
2 + 2-signature emerges in several physical context, including self-dual gravity a la
35 Plebanski (see Ref. 6 and references therein), consistent N = 2 superstring theory
as discussed by Ooguri and Vafa,7,8 N = (2, 1) heterotic string.9–12 Moreover, it has
37 been emphasized13,14 that Majorana–Weyl spinor exists in space–time of 2+2 signa-
ture. Even cosmologically there is a wisdom15 that the 2+2 signature is interesting.

1
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

2 C. Castro & J. A. Nieto

1 In Refs. 23–26 it was shown how a N = 2 supersymmetric Wess–Zumino–

Novikov–Witten model valued in the area-preserving (super)diffeomorphisms group
3 is self-dual supergravity in 2 + 2 and 3 + 1 dimensions depending on the signatures
of the base manifold and target space. The interplay among W∞ gravity, N = 2
5 strings, self-dual membranes, SU(∞) Toda lattices and SU∗ (∞) Yang–Mills instan-
tons in 2 + 2 dimensions can be found also in Refs. 23–26.
7 More recently, using the requirement of the SL(2, R) and Lorentz symmetries it
has been proved16 that 2 + 2 target space–time of a 0-brane is an exceptional signa-
9 ture. Moreover, following an alternative idea to the notion of worldsheets for world-
sheets proposed by Green17 or the 0-branes condensation suggested by Townsend18
11 it was also proved in Ref. 16 that special kind of 0-brane called quatl19,20 leads
to the result that the 2 + 2-target space–time can be understood either as 2 + 2-
13 worldvolume space–time or as 1 + 1 matrix-brane.
Another recent motivation for a physical interest in the 2 + 2 signature has
15 emerged via Duff’s21 discovery of hidden symmetries of the Nambu–Goto action. In
fact, this author was able to rewrite the Nambu–Goto action in a 2+2 target space–
17 time in terms of a hyperdeterminant, reveling apparently new hidden symmetries
of such an action. More recently the Duff’s observation has been linked with the
19 matrix-brane idea.22
Considering seriously the possibility that the (2 + 2)-dimensional “space–time”
21 is an exceptional signature one may wonder what is the connection between
(2 + 2)-dimensional “space–time” and other exceptional structures in physics such
23 as black-holes. In this respect it becomes convenient to discuss black-holes physics
from modern perspective. In particular, it become convenient to clarify the many
25 subtleties behind the introduction of a true point-mass source at r = 039 and the
admissible family of radial functions R(r) in the static spherically symmetric solu-
27 tions of Einstein field equations29–38 (see also Refs. 42–45).
We begin by writing down the class of static spherically symmetric (SSS) vacuum
solutions of Einstein’s equations46 studied by Refs. 29–32, 52–58, 42–45 and 72,
among many others, given by a infinite family of solutions parametrized by a family
of admissible radial functions R(r) (in c = 1 units)
   −1
2 2GN Mo 2 2GN Mo
(ds) = 1 − (dt) − 1 − (dR)2 − R2 (r)(dΩ)2 , (1.1)
R R
where the solid angle infinitesimal element is
(dΩ)2 = (dφ)2 + sin2 (φ)(dθ)2 . (1.2)
This expression of the metric in terms of the radial function R(r) (a radial gauge)
29 does not violate Birkoff’s theorem since the metric (1.1), (1.2) expressed in terms
of the radial function R(r) has exactly the same functional form as that required
31 by Birkoff’s theorem and 0 ≤ r ≤ ∞. Metrics of the form (1.1) were employed
by Ref. 94 based on the nonperturbative renormalization group flow and running
33 Newtonian coupling G = G(r) in quantum Einstein gravity.90–93
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

On (2 + 2)-Dimensional Space–Times, Strings and Black Holes 3

1 There are two interesting cases to study based on the boundary conditions
obeyed by R(r): (i) the Hilbert textbook (black hole) solution49–51 based on the
3 choice R(r) = r obeying R(r = 0) = 0, R(r → ∞) → r. And (ii) the controversial
(erroneous) Abrams–Schwarzschild radial gauge based on choosing the cutoff R(r =
5 0) = 2GN M such that gtt (r = 0) = 0 which apparently seems to “eliminate” the
horizon and R(r → ∞) → r. This was the original solution of 1916 found by
7 Schwarzschild. However, the choice R(r = 0) = 2GM has a serious flaw and is: how
is it possible for a point-mass at r = 0 to have a nonzero area 4π(2G N M )2 and a
9 zero volume simultaneously? so it seems that one is forced to choose the Hilbert
gauge R(r = 0) = 0 and retain only those metrics that are diffeomorphic to the
11 Hilbert textbook black hole solution only.
Nevertheless there is a very specific radial function (never studied before to our
13 knowledge) R(r) = r+2GN M Θ(r) 36 that yields a metric which is not diffeomorphic
to the Hilbert textbook solution based on the Heaviside step functiona which is
15 defined Θ(r) = 1 when r > 0, Θ(r) = −1 when r < 0 and Θ(r = 0) = 0 (the
arithmetic mean of the values at r > 0 and r < 0). The Heaviside step function
17 behavior at r = 0 given by Θ(r = 0) = 0 will ensure us that now we can satisfy
the required condition R(r = 0) = r = 0, consistent with our intuitive notion that
19 the
p spatial area and spatial volume of a point r = 0 has to be zero. Since r =
± x2 + y 2 + z 2 , a negative r branch is mathematically possible and is compatible
21 with the double covering inherent in the Fronsdal–Kruskal–Szekeres60–62 analytical
continuation in terms of the u, v coordinates. Each point of space–time inside
23 r < 2GN M is represented twice (black hole and white hole picture). However there
is a fundamental difference (besides others) with the Fronsdal–Kruskal–Szekeres
25 extension into the interior of r = 2GM , their metric description is no longer static
in r < 2GM , whereas in our case the metric is static for all values of r.
27 Thus the scalar curvature associated to the point mass delta function source
−2GN M δ(r)/R2 (dR/dr) 39 does not always remain invariant of the radial gauge
29 chosen. In the very special case chosen by Schwarzschild in 1916 given by R 3 =
r3 + (2GN M )3 the scalar curvature and measure remains the same as in the Hilbert
31 textbook choice R(r) = r due to the relation R 2 dR = r2 dr. But this was a his-
torical fluke. An unfortunate accident which has impeded the progress for 90 years
33 because many were misled into thinking that any radial gauge choice was always
equivalent to a naive radial reparametrization r → r 0 of the Hilbert metric. It is not
35 because having a family of nondiffeormorphic metrics, parametrized by a family of
inequivalent radial gauges belonging to different gauge orbits, is not the same thing
37 as having a family of naive radial changes of coodinates r → r 0 associated to a fixed
and given fiduciary metric.
39 The reason why there are metrics which are not diffeomorphic to the Hilbert
textbook solution is due to the fact that there are orbits obtained by exponentiation

a We thank Michael Ibison for pointing out the importance of the Heaviside step function and the
use of the modulus |r| to account for point mass sources at r = 0.
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

4 C. Castro & J. A. Nieto

1 of generators of diffeomorphisms that yield diffeomorphisms which are not connected

to the identity and which still may act trivially at infinity (Marsden theorem). The
3 identity element of the diffs group is in our case related to the Hilbert textbook
trivial radial gauge-function R(r) = r. Consequently, there are radial gauges which
5 are not obtained by a naive radial reparametrization r → r 0 of the Hilbert textbook
metric and correspond to metrics which are not physically equivalent to it. More-
7 over, Donaldson showed that in D = 4 one has an infinite number of inequivalent
differential structures, i.e. manifolds that are homeomorphic (topologically equiva-
9 lent) but are not diffeomorphic. The presence of matter (singularity) at r = 0 and
the different choices of inequivalent radial gauges should single out the particular
11 differential structure in D = 4.
There is an essential technical subtlety required in order to generate δ(r) terms
13 in the right-hand side of Einstein’s equations. One must replace everywhere r → |r|
as required when point-mass sources are inserted. A rigorous mathematical treat-
15 ment of Colombeau’s theory of nonlinear distributions can be found in Refs. 63–66.
The Newtonian gravitational potential due to a point-mass source at r = 0 is given
17 by −GN M/|r| and is consistent with Poisson’s law which states that the Laplacian
of the Newtonian potential −GM/|r| is 4πGρ where ρ = (M/4πr 2 )δ(r) in Newto-
19 nian gravity. However, the Laplacian in spherical coordinates of (1/r) is zero. For
this reason, there is a fundamental difference in dealing with expressions involving
21 absolute values |r| like 1/|r| from those which depend on r like 1/r.59 Therefore
the radial gauge must be chosen by R(|r|) = |r| + 2GN M Θ(|r|). Had one not use
23 |r| in the expression for the metric, one will not generate the desired δ(r) terms in
the right-hand side of Einstein’s equations Rµν − 21 gµν R = −8πGN Tµν 6= 0, and
25 one would get an expression identically equal to zero (consistent with the vacuum
solutions in the absence of matter) instead of the δ(r) terms.39
27 To sum up, by using R(|r|) = |r| + 2GN M Θ(|r|), we safely have that R(|r|) =
|r| + 2GN M , when r > 0 and the horizon can the be displaced from r = 2GN M
29 to a location as arbitrarily close to r = 0 as desired rHorizon → 0. To be more
precise, the horizon actually never forms since at r = 0 one hits the singularity.
31 Also, R ∼ r for r  2GN M and one recovers the correct Newtonian limit in the
asymptotic regime. It is now, via the Heaviside step function, that we may maintain
33 the correct behavior R(|r|) = |r|, when r = 0, and such that we can satisfy the
required condition R(r = 0) = r = 0, consistent with our intuitive notion that
35 the spatial area and spatial volume of a point r = 0 has to be zero. The metric
is smooth and differentiable for all r > 0 and one will have Rµν = R = 0 (in the
37 region r > 0 empty of matter and radiation). The metric is discontinuous only at
the location of the point mass singularity r = 0 whose worldline which may be
39 thought of as the boundary of space–time. The scalar curvature is infinite at r = 0
due to the delta function point mass source at r = 0, it jumps from zero to infinity
41 at r = 0.
And most importantly, a radial reparametrization r → r 0 (r) leaves invariant
the scalar curvature and the measures associated with a given choice of the radial
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

On (2 + 2)-Dimensional Space–Times, Strings and Black Holes 5

function R1 (r):
4πR12 (r)dR1 (r)dt = 4πR102 (r0 )dR10 (r0 )dt , (1.3a)
2GN M
R1 (r) = − 2 δ(r) = R01 (r0 )
R1 (r)(dR1 /dr)
2GN M
=− δ(r0 ) . (1.4a)
R102 (r0 )(dR10 (r0 )/dr0 )
Choosing a different radial function R2 (r) gives under a radial reparametrization
r → r0 (r):
4πR22 (r)dR2 (r)dt = 4πR202 (r0 )dR20 (r0 )dt , (1.3b)
2GN M
R2 (r) = − δ(r) = R02 (r0 )
R22 (r)(dR2 /dr)
2GN M
=− δ(r0 ) . (1.4b)
R202 (r0 )(dR20 (r0 )/dr0 )
1
In the same manner that one must not confuse active and passive diffeomor-
phisms we have
2GN M 2GN M
R(r) 6= r0 (r) ⇒ R(r) = − 2 δ(r) = − 2 δ(R(r))
R (dR/dr) R (r)
2GN M 2GN M
6= − δ(r) = − 02 δ(r0 (r)) . (1.5)
r02 (r)(dr0 /dr) r (r)
Because the scalar curvature is an explicit function of the radial function R(r)
3 given by this expression: −2GM δ(r)/R2 (r)(dR/dr) = −2GM δ(R(r))/R2 (r) we
can see that the scalar curvature does not remain invariant of the infinite number
5 of possible choices of the radial functions R(r), except in the anomalous case when
R3 = r3 + (2GM )3 (the radial gauge chosen by Schwarzschild in 1916) that leads to
7 −2GM δ(r)/r 2 , and which accidentally happens to agree with the scalar curvature
in the Hilbert gauge R(r) = r.
What remains invariant of the choices R(r) is the action
ZZ  
1 2GN Mo
S=− − 2 δ(r) (4πR2 dR dt)
16πGN R (dR/dr)
ZZ  
1 2GN Mo
=− − δ(r) (4πr2 dr dt) . (1.6)
16πGN r2
The Euclideanized Einstein–Hilbert action associated with the scalar curvature
delta function is obtained after a compactification of the temporal direction along a
circle S 1 giving an Euclidean time coordinate interval of 2πtE and which is defined
in terms of the Hawking temperature TH and Boltzman constant kB as 2πtE =
(1/kB TH ) = 8πGN Mo .
4π(GN Mo )2 4π(2GN Mo )2 Area
SE = 2 = = . (1.7)
LPlanck 4L2Planck 4L2Planck
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

6 C. Castro & J. A. Nieto

1 It is interesting that the Euclidean action SE (in ~ units) is precisely the same as
the black hole entropy S in Planck area units. This result holds in any dimensions
3 D ≥ 3. This is not a numerical coincidence. Furthermore, the action is invariant of
the choices of R(r), whether or not it is the Hilbert textbook choice R(r) = r or
5 another. The choice of the radial function R(r) amounts to a radial gauge that leaves
the action invariant but it does not leave the scalar curvature, nor the measure of
7 integration, invariant. Only the action (integral of the scalar curvature) remains
invariant.
9 The action–entropy connection has been obtained from a different argument,
for example, by Padmanabhan40 by showing how it is the surface term added to
11 the action which is related to the entropy, interpreting the horizon as a boundary
of space–time. The surface term is given in terms of the trace of the extrinsic cur-
13 vature of the boundary. The surface term in the action is directly related to the
observer-dependent-horizon entropy, such that its variation, when the horizon is
15 moved infinitesimally, is equivalent to the change of entropy dS due to the vir-
tual work. The variational principle is equivalent to the thermodynamic identity
17 T dS = dE + p dV due to the variation of the matter terms in the right-hand side.
A bulk and boundary stress energy tensors are required to capture the Hawking
19 thermal radiation flux seen by an asymptotic observer at infinity as the black hole
evaporates.
21 With these modern developments at hand one may proceed to find “black-hole”
type solutions of the Einstein field equations for a (2 + 2)-dimensional “space–
23 time.” In Sec. 2 we present static hyperbolic solutions in a (2 + 2)-dimensional
“space–time” and describe its differences with the corresponding solution in 3 + 1
25 dimensions. In Secs. 3 and 4, we present the straightforward computations of the
static circular symmetric solutions of Einstein field equations in 2 + 2 dimensions.
27 Finally, in Sec. 5 we show how the 1 + 1 Bars–Witten stringy black-hole solution
can be embedded into the (3 + 1)-dimensional solution of the appendix and discuss
29 the “stringy” nature behind a point-mass. Black holes in a (2 + 2)-dimensional
“space–time” from the perspective of complex gravity in 1 + 1 complex dimensions
31 and its quaternionic and octonionic gravity extensions deserve furher investigation.
In the appendix we construct Schwarzschild-like solutions in dimensions D ≥ 4.

33 2. Static Hyperbolic Symmetric Solution in 2 + 2 Dimensions

Consider the vacuum static spherically symmetric solutions of Einstein field equa-
tions in a space–time of (3 + 1)-signature
Rµν = 0 (2.1)
of the form
ds2 = −eµ(r) (dt1 )2 + eα(r) dr2 + R2 (r)dΩ2 , (2.2)
where
dΩ2 = dφ2 + sin2 φ dθ2 . (2.3)
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

On (2 + 2)-Dimensional Space–Times, Strings and Black Holes 7

The solutions are

(dR/dr)2
 
2 α
ds = − 1 − (dt1 )2 + dr2 + R2 (r)dΩ2 , (2.4)
R (1 − α/R)
where α is a parameter that has mass dimensions. Several remarks are now in order
pertaining whether or not a Wick rotation of the metric (2.4) furnishes solutions
to the vacuum field equations for the signature 2 + 2. A naive Wick rotation of the
angle coordinate φ → iφ = χ in the above solutions (2.4) yields
sin2 (φ) → sin2 (iφ) = − sinh2 (χ) , dφ2 → −dχ2 , (2.5)
1 and due to the two sign changes in (2.5) one would have a 1 + 3 signature instead
of a split 2 + 2 signature.
A Wick rotation of θ → iθ = χ, (dθ)2 → −(dχ)2 yields a 2 + 2 signature but
since the range of the only remaining angle φ is [0, π], instead of [0, 2π], and one
will no longer cover the space completely. Furthermore, since there is a signature
change (a sign change in one of the metric components gθθ ) the connection and
curvature expressions will be modified accordingly and there is no reason now why
the vacuum field equations should be satisfied. In the next section we will find
explicit solutions in the static circular symmetric case:
ds2 = −eµ̃(R(ρ)) (dt1 )2 − eν̃(R(ρ)) (dt2 )2 + eα̃(R(ρ)) (dR(ρ))2 + (R(ρ))2 dθ2 ,
3 where the rho function R(ρ) is now a function of ρ, the radius of a circle ρ2 = x2 +y 2 .
In order to construct solutions with topology H 3 × R where H3 is a three-
dimensional pseudosphere (a hyperboloid) of radius R parametrized by the coordi-
nates ψ, θ, χ as
x = R cosh χ cos θ , y = R cosh χ sin θ ,
(2.6)
t1 = R sinh χ cos ψ , t2 = R sinh χ sin ψ ,
where −∞ ≤ χ ≤ ∞ and 0 ≤ θ ≤ 2π; 0 ≤ ψ ≤ 2π such that the flat space–time
metric in 2 + 2 dimensions is
ds2 = −(dt1 )2 − (dt2 )2 + (dx)2 + (dy)2
= (dR)2 + R2 [cosh2 χ(dθ)2 − sinh2 χ(dψ)2 − (dχ)2 ] . (2.7a)
From Eq. (2.6) we infer that the three-dimensional pseudosphere H 3 is repre-
sented analytically by
−(t1 )2 − (t2 )2 + x2 + y 2 = R2 . (2.7b)
The curved space–time metric we are interested involve the two functions Σ =
Σ(R) and f˜ = f(Σ(R))
˜ = f (R) such that
˜
ds2 = ef (Σ) (dΣ)2 + Σ2 [cosh2 χ(dθ)2 − sinh2 χ(dψ)2 − (dχ)2 ]
 2
dΣ
= ef (R) (dR)2 + Σ2 (R)[cosh2 χ(dθ)2 − sinh2 χ(dψ)2 − (dχ)2 ]
dR
= eµ(R) (dR)2 + Σ2 (R)[cosh2 χ(dθ)2 − sinh2 χ(dψ)2 − (dχ)2 ] , (2.8)
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

8 C. Castro & J. A. Nieto

1 where we have defined eµ(R) ≡ ef (R) (dΣ/dR)2 . The flat space–time metric (2.7) is
recovered from (2.8) in the limit R → ∞ such that µ(R) → 0 and Σ(R) ∼ R.
Another interesting parametrization r ≥ 0, and −∞ ≤ ξ ≤ ∞; 0 ≤ θ ≤ 2π is
t2 = r sinh ξ , x = r cosh ξ cos θ , y = r cosh ξ sin θ , (2.9)
where r is the throat size of the two-dimensional hyperboloid H 2 defined in terms
of t2 , x, y as
−(t2 )2 + x2 + y 2 = r2 (2.10)
and the flat space–time metric −(dt1 )2 − (dt2 )2 + (dx)2 + (dy)2 can be recast as
ds2 = −(dt1 )2 + (dr)2 + r2 [cosh2 ξ(dθ)2 − (dξ)2 ] . (2.11)
3 Notice that we have a 2 + 2 signature in Eq. (2.11), as one should, and that there is
a difference between the forms of the metric in Eqs. (2.7) and (2.11). The topology
5 corresponding to Eq. (2.7) is H3 × R∗ where H3 is a three-dimensional hyperboloid
(a three-dimensional pseudosphere); whereas, instead, the topology corresponding
7 to Eq. (2.11) is R × R∗ × H2 .
R∗ is the half-interval [0, ∞] representing the values of the radial coordinates.
9 In Eq. (2.7) the three-dimensional hyperboloid (pseudosphere) of fixed radius R
is spanned by the three coordinates θ, ψ, χ as indicated by Eq. (2.6). Whereas
11 in Eq. (2.11), one temporal variable t1 is characterized by the real line R and
whose values range from −∞, +∞, and the other temporal variable t2 is one of the
13 three coordinates (t2 , x, y) which parametrized the two-dimensional hyperboloid H 2
described by Eq. (2.10).
A curved space–time version of Eq. (2.11) is
ds2 = −eµ(r) (dt1 )2 + eν(r) (dr)2 + (R(r))2 [cosh2 ξ(dθ)2 − (dξ)2 ] . (2.12a)
15 The metric in Eq. (2.12a) whose signature is 2 + 2 is the hyperbolic version of
the Schwarzschild metric. One can replace r → R(r) since Einstein’s equations do
17 not determine the form of the radial function R(r) as explained in the appendix.
The global topology of the solutions depends on the choices of R(r). We still must
19 determine what are the functional forms of µ(r) and ν(r). In order to go from
the solid angle (dΩ)2 = sin2 (φ)(dθ)2 + (dφ)2 to cosh2 ξ(dθ)2 − (dξ)2 one must first
21 perform the change of coordinates φ → π/2 + φ such that sin2 φ → cos2 (φ) and
then Wick rotate φ → φ = iξ so that cos2 (φ) → cosh2 ξ and (dφ)2 = −(dξ)2 .
In the appendix we find the solutions to Einstein’s vacuum field equations in D
dimensions for metrics whose signature is (D − 2) + 2 (two times) associated with
a (D − 2)-dimensional homogeneous space of constant positive (negative) scalar
curvature. In particular when D = 4 and the two-dimensional homogeneous space
H2 has a constant positive scalar curvature, like two-dimensional de Sitter space,
the metric components, in natural units G = ~ = c = 1, are given by
(dR/dr)2
 
βM
g t1 t1 = − 1 − , grr = , β = const (2.12b)
R(r) (1 − βM/R(r))
23
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

On (2 + 2)-Dimensional Space–Times, Strings and Black Holes 9

1 which are almost identical to the components appearing in the Schwarzchild solu-
tions for signature 3 + 1. The two-dimensional hyperboloid defined by Eq. (2.10)
3 coincides with a two-dimensional de Sitter space of constant positive scalar curva-
ture. Anti-de Sitter space has a constant negative scalar curvature.
There is a physical singularity at r = 0, the location of the point mass source,
when the hyperboloid H2 degenerates to a cone since the throat size r has been
pinched to zero. When the radial function is chosen to be R 3 = r3 +(βM )3 ⇒ R(r =
0) = βM then grr (r = 0) = ∞ and gt1 t1 (r = 0) = 0. The proper circumference for
this choice R3 = r3 + (βM )3 is
C(r, ξ) = 2πR(r) cosh ξ 0 ⇒ C(r = 0, ξ) = 2πβM cosh ξ . (2.13)
The proper area for a given value of r is
Z +∞
A(r) = 2πR2 (r) cosh ξ dξ = 2πR2 (r)2 sinh ξ → ∞ (2.14)
−∞

5 and diverges as ξ → ∞ because the two-dimensional hyperboloid is not compact.

If one chooses R(r) = r, then R(r = 0) = 0, so the proper circumference is zero
7 (for finite ξ) and the proper area corresponding to r = 0 is 0 × ∞ = ∞ since sinh ξ
approaches infinity faster than r 2 approaches zero.
Another parametrization is
t2 = r cosh ξ , x = r sinh ξ cos θ , y = r sinh ξ sin θ , (2.15)
where the thoat size r is defined in terms of t2 , x, y as
−(t2 )2 + x2 + y 2 = −r2 (2.16)
which can be obtained from Eq. (2.10) by r 2 → −r2 . Equation (2.16) represents
analytically the two disconnected branches of a two-dimensional hyperboloid:
ds2 = −(dt1 )2 − (dt2 )2 + (dx)2 + (dy)2
= −(dt1 )2 − (dr)2 + r2 [sinh2 ξ(dθ)2 + (dξ)2 ] . (2.17)
9 Notice the sign change −dr 2 in Eq. (2.15) as one must have if one persists in having
a 2 + 2 signature. In this case the coordinate r must be interpreted as a “radial
11 time.”
The curved space–time version of (2.17) would be
ds2 = −eα(r) (dt1 )2 − eβ(r) (dr)2 + (R(r))2 [sinh2 ξ(dθ)2 + (dξ)2 ] , (2.18)
where α(r) and β(r) are two functions to be determined by solving Einstein’s equa-
13 tions. The functional form of α(r), β(r) dif f ers from the functions µ(r), ν(r) in
Eqs. (2.12a) and (2.12b) due to a crucial sign change in the grr component of the
15 metric in Eq. (2.18).
Concluding, we have 3 interesting cases described by the metrics of 2 + 2 signa-
17 ture given by Eqs. (2.8), (2.12) and (2.18). The 2 + 2 hyperbolic-symmetric version
of Schwarzschild’s 3 + 1 solution is given by Eqs. (2.12a) and (2.12b).
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

10 C. Castro & J. A. Nieto

1 3. Static Circular Symmetric Solution in 2 + 2 Dimensions

Let us look for a solution of the field equations of the form
ds2 = −eµ̃(R) (dt1 )2 − eν̃(R) (dt2 )2 + eα̃(R) dR2 + R2 dθ2
= −eµ(ρ) (dt1 )2 − eν(ρ) (dt2 )2 + eα(ρ) dρ2 + R2 (ρ)dθ2 , (3.1a)
where
 2
dR
µ̃(R(ρ)) = µ(ρ) , ν̃(R(ρ)) = ν(ρ) , eα̃(R(ρ)) = eα(ρ) . (3.1b)
dρ
The only nonvanishing Christoffel symbols are
1 0 1 0 R0 1 0 µ−α
Γ131 = µ , Γ232 = ν , Γ434 = , Γ311 = µe ,
2 2 R 2
(3.2)
1 1
Γ322
= ν 0 eν−α , Γ344 = −e−α RR0 , Γ333 = α0 ,
2 2
and the only nonvanishing Riemann tensor are
1 1
R1212 = µ0 ν 0 eν−α , R1414 = − µ0 e−α RR0 ,
4 2
1 0 0 µ−α 1
R2121 = µνe , R2424 = − ν 0 e−α RR0 ,
4 2
1 0 µ−α R0 1 0 ν−α R0
R4141 = µe , R4242 = νe ,
2 R 2 R
(3.3)
1 1 1 1 1 1
R1313 = − µ00 − µ02 + α0 µ0 , R2323 = − ν 00 − ν 02 + α0 ν 0 ,
2 4 4 2 4 4
00 0
 
R 1 0R µ−α 1 00 1 02 1 0 0
R4343 =− + α , 3
R131 = e µ + µ − αµ ,
R 2 R 2 4 4
   
1 00 1 02 1 0 0 1
R3232 = eν−α ν + ν − α ν , R3434 = e−α R α0 R0 − R00 .
2 4 4 2
The field equations are
 
1 00 1 02 1 0 0 1 0 0 1 0 R0
R11 = eµ−α µ + µ + µν − αµ + µ = 0, (3.4)
2 4 4 4 2 R
 
1 00 1 02 1 0 0 1 0 0 1 0 R0
R22 = eν−α ν + ν + µν − αν + ν = 0, (3.5)
2 4 4 4 2 R
1 1 1 1 1
R33 = − µ00 − µ02 + µ0 α0 − ν 00 − ν 02
2 4 4 2 4
1 1 R0 R00
+ α0 ν 0 + α0 − = 0, (3.6)
4 2 R R
and
 
1 1 1
R44 = e−α R − µ0 R0 − ν 0 R0 + α0 R0 − R00 = 0 . (3.7)
2 2 2
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

On (2 + 2)-Dimensional Space–Times, Strings and Black Holes 11

From (3.7) we get

2R00
α0 = µ 0 + ν 0 + . (3.8)
R0
Substituting (3.8) into (3.4) and (3.5) we obtain
 00 
µ00 R R0
= − (3.9)
µ0 R0 R
and
 
ν 00 R00 R0
= − , (3.10)
ν0 R0 R
respectively. Equations (3.9) and (3.10) can be integrated to give
R0
µ0 = a (3.11)
R
and
R0
ν0 = b , (3.12)
R
respectively, where a and b are constants. Substituting (3.11) and (3.12) into (3.8)
leads to
R0 R0 2R00
α0 = a +b + 0 . (3.13)
R R R
The expressions (3.11)–(3.13) can be solved. We get
µ = a ln R/c , (3.14)

ν = b ln R/d (3.15)
and
α = a ln R/c + b ln R/d + 2 ln R0 + f , (3.16)
where c, d and f are arbitrary constants. If we substitute (3.14)–(3.16) into (3.6)
we find
   0  0 
1 R00 R02 1 2 R02 1 R R R0 2R00
− a − 2 − a + a a +b + 0
2 R R 4 R2 4 R R R R
 00    
1 R R02 1 2 R02 1 R0 R0 R0 2R00
− b − 2 − b + b a +b + 0
2 R R 4 R2 4 R R R R
 0 
1 R R0 2R00 R0 R00
+ a +b + 0 − = 0. (3.17)
2 R R R R R
This can be reduced to
  02
1 R
a + ab + b = 0. (3.18)
2 R2
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

12 C. Castro & J. A. Nieto

Excluding the solutions

R = const (3.19)

Eq. (3.18) gives

1
a + ab + b = 0 . (3.20)
2
Therefore we have shown why the form of R = R(ρ) can be completely arbitrary
while one must have the following constraint among the constants:
2a
b=− , (3.21)
(a + 2)
1 where we assumed that a + 2 6= 0.
A trivial solution of Eq. (3.20) is a = b = 0 which leads to µ = ν = 0 and
α = 2 ln(dR/dρ), when f = 0, yielding the metric

ds2 = −(dt1 )2 − (dt2 )2 + dR(ρ)2 + R2 (ρ)dθ2 , (3.22)

the flat space–time metric is attained when R(ρ) = ρ, and also for any function
3 R(ρ) with the asymptotic property such that for very large values of ρ it behaves
R ∼ ρ.

5 4. An Explicit Nontrivial Solution

We have seen that the trivial flat space–time solutions (3.22) are obtained when
a = b = f = 0 and when R(ρ) = ρ. In order to find interesting nontrivial solutions
we should have a nontrivial rho function R(ρ) 6= ρ. Let us consider two particular
cases of (3.21). In the first case taking a = 2 from Eq. (3.21) we get b = −1.
Similarly, in the second case by setting a = −1 in Eq. (3.21) implies b = 2. Thus in
the first case (3.14)–(3.16) become

µ = 2 ln R/c , (4.1)
ν = − ln R/d (4.2)

and

α = 2 ln R/c − ln R/d + 2 ln R0 + f . (4.3)

While in the second case we find

µ = − ln R/c , (4.4)
ν = 2 ln R/d (4.5)

and

α = − ln R/c + 2 ln R/d + 2 ln R0 + f . (4.6)

March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

On (2 + 2)-Dimensional Space–Times, Strings and Black Holes 13

An interesting possibility arises by setting c = d = M and f = 0. In the first case we

get that the metric in 2 + 2 dimensions ends up being expressed in the R-variable as
    
R 2 M R
ds2 = − ) (dt1 )2 − (dt2 )2 + (dR)2 + R2 (dθ)2 , (4.7)
M R M
while in the second case we obtain
   2  
2 M 2 R 2 R
ds = − (dt1 ) − (dt2 ) + (dR)2 + R2 (dθ)2 . (4.8)
R M M
1 Notice that in both solutions (4.7) and (4.8) there is a kind of duality in the two
times t1 and t2 factors.
Equations (4.7) and (4.8) can be written as
2 2
   
2 (dθ) − (dt1 )
 
2 R 2 R 2
ds = − (dt2 ) + (dR) + R , (4.9a)
M M M2
2 2
   
2 (dθ) − (dt2 )
 
2 R 2 R 2
ds = − (dt1 ) + (dR) + R . (4.9b)
M M M2
3 As announced earlier, the form of the rho function R(ρ) is undetermined. Any
arbitrary choice of R(ρ) solves Einstein’s equations.
A study reveals that a rho function R(ρ) given by
1 1 1
= + , (4.10)
R ρ M
in units of G = ~ = c = 1 is an appropriate choice. When ρ = 0, R = 0 and when
ρ = ∞ we have R(ρ = ∞) = M , so we do recover an asymptotically flat space–time
metric at spatial ρ = ∞ given by
ds2 = −(dt1 )2 − (dt2 )2 + (dR)2 + R2 (dθ)2 = −(dt1 )2 − (dt2 )2 + M 2 (dθ)2 . (4.11)
5 Asymptotic infinity is defined by the condition R(ρ = ∞) = M . It is the three-
dimensional asymptotic boundary of the (2+2)-space–time. It is a three-dimensional
7 manifold of topology S 1 × R2 . The radius of S 1 is R = M . When ρ = 0 we have in
Eq. (4.7) that R(ρ = 0) = 0, so the metric component g22 (ρ = 0) = ∞ and there is
9 a metric singularity at ρ = 0 as expected. Conversely, in Eq. (4.8) the singularity
occurs in the component g11 (ρ = 0) = ∞, instead.

11 5. Stringy 1 + 1 Black Holes Embedded in 3 + 1 and

2 + 2 Dimensions
13 One of the main topics of the present work has been to link the 2 + 2 signature with
the black hole concept, i.e. space–times with singularities. We have shown that there
15 are many different interesting ways to do this. In Sec. 2 we presented three very
diferent cases associated with hyperboloids. In particular, in the static hyperbolic-
17 symmetric version of the Schwarschild case given by Eqs. (2.12a) and (2.12b), there
is singularity at r = 0 which is associated with the conical geometry resulting from
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

14 C. Castro & J. A. Nieto

1 having pinched to zero size r = 0 the throat of the hyperboloid H 2 and which
is quite different from the spherically symmetric case in 3 + 1 dimensions. In the
3 static circular symmetric case developed in Secs. 3 and 4 we obtained solutions with
singularties at ρ = 0 and whose asymptotic ρ → ∞ limit leads to a flat (1 + 2)-
5 dimensional boundary of topology S 1 ×R2 where the radius of S 1 is R(ρ = ∞) = M .
One further interesting possibility may arise if we split the 2 + 2 metric as the
diagonal sum of two 1 + 1 metrics in the form
ds2 = gab (x)dxa dxb + gmn (y)dy m dy n , a, b = 1, 2 , m, n = 3, 4 . (5.1)
In this case one may look for solutions like
du dv dw dz
ds2 = + , (5.2)
1 − uv 1 − wz
where we have set the value of the mass parameter 2M = 1. Such mass parameter is
7 required on physical grounds and also because the denominators in Eq. (5.2) must
be dimensionless.
The metric of Eq. (5.2) can be understood as the diagonal sum of two 1 + 1
black holes solutions95–97 and whose singularities are located at uv = 1 and wz = 1
respectively. There are two horizons. The region outside the first horizon is indicated
by u ≥ 0 ≥ v and v ≥ 0 ≥ u; and the region inside the first horizon is indicated
by 1 ≥ uv ≥ 0 and u, v ≥ 0. Similar considerations apply to the second horizon by
exchanging u ↔ w and v ↔ z. The lightcone coordinates are defined by
1
u = exp[x + t1 + log(1 − e−2x )] = X + T1 ,
2
(5.3a)
1
v = − exp[x − t1 + log(1 − e −2x
)] = X − T1 ,
2
1
w = exp[y + t2 + log(1 − e−2y )] = Y + T2 ,
2
(5.3b)
1
z = − exp[y − t2 + log(1 − e )] = Y − T2 .
−2y
2
Conformally flat solutions of the form
ds2 = eΥ(x,y,t1,t2 ) [(dx)2 − (dt1 )2 + (dy)2 − (dt2 )2 ] , (5.4)
9 where Υ(x, y, t1 , t2 ) has a similar singularity structure as the metric in Eq. (5.2)
are worth exploring also.
The Bars–Witten black hole (1 + 1)-dimensional metric (setting 2M = 1) is
du dv
ds2 = (dr)2 − tanh2 (r)(dt)2 = − (5.5)
1 − uv
with
1
u= exp[r + t + log(1 − e−2r )] ,
2
(5.6)
1
v = − exp[r − t + log(1 − e−2r )] .
2
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

On (2 + 2)-Dimensional Space–Times, Strings and Black Holes 15

1 The Euclidean analytical continuation of the metric in Eq. (5.5) is obtained by

setting θ = it, such that the metric is ds2 = dr2 + tanh2 r dθ2 and its Euclidean
3 geometry has the shape of a semiinfinite cigar that asymptotically approaches R 1 ×
S 1 for r → ∞. We should notice that the Lorentzian metric of Eq. (5.5) has a
5 singularity at a complex value r = 0 + iπ/2 (setting 2M = 1) since tanh2 (iπ/2) =
− tan2 (π/2) = −∞ which is consistent with the singularities at the location where
7 uv = − 41 e2r (1−e−2r )2 = 1, when r = 0+iπ/2, and a horizon at r = 0, since uv = 0
when r = 0.
However this is not the end of the story. The Bars–Witten black hole in (1 + 1)-
dimensional is obtained from a gauged Sl(2, R)/U(1) WZNW model with central
charge c = 2 + 6/k and is a consistent bosonic string background solution in a 1 + 1
target background given by the two-dimensional coset Sl(2, R)/U(1). Namely, the
CFT corresponding to the gauged Sl(2, R)/U(1) WZNW model with central charge
c = 2 + 6/k is a solution of equations derived from the vanishing beta functions
required by conformal invariance of the nonlinear sigma model. For example, the
relevant massless bosonic closed-string fields in a (D = 26)-dimensional target back-
ground (a different CFT) are the antisymmetric tensor Bµν (X ρ (σ a )); the dilaton
Φ(X ρ (σ a )) and the gravitational field gµν (X ρ (σ a ))); where σ a = σ 1 , σ 2 are the
worldsheet variables. The conditions for the vanishing of the one loop beta func-
tions, required by Weyl invariance of the nonlinear sigma model, to leading order
in the string tension α0 turn out to be99
1
Rµν + Hµλρ Hνλρ − 2Dµ Dν Φ = 0 , (5.7a)
4
λ λ
Dλ Hµν − 2(Dλ Φ)Hµν = 0, (5.7b)
1
4(Dµ Φ)2 − 4Dµ Dµ Φ + R + Hµνρ H µνρ = 0 , (5.7c)
12
where
Hµνρ = ∂µ Bνρ + ∂ρ Bµν + ∂ν Bρµ , (5.7d)
9 is the third rank antisymmetric tensor field strength that is invariant under the
transformations δBµν = ∂µ Λν − ∂ν Λµ . For details of quantum nonlinear sigma
11 models, conformal field theory, supersymmetry, black holes and strings we refer to
the monograph by Ketov.98
13 The only consistent (2+2)-dimensional gravitational backgrounds on which N =
2 strings7,8 (strings with worldsheet supersymmetry) can propagate are those that
15 are self-dual and which solve the Plebanski heavenly equations in 2 + 2 dimensions.
Self dual gravitational backgrounds in four dimensions are Ricci flat whose metric
17 is given in terms of a Kahler potential. However, the metric in Eq. (5.2) is not
Ricci flat since the (1 + 1)-dimensional black hole metric is not Ricci flat. Such
19 metric in Eq. (5.5) is not a solution of the vacuum Einstein field equations, it is
a solution of Eqs. (5.7) (without Kalb–Ramond fields Bµν ) where the role of the
21 dilaton Φ = ln(1 − uv) is essential.
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

16 C. Castro & J. A. Nieto

Nevertheless, we will show how the Bars–Witten (1 + 1)-dimensional black hole

metric can be embedded into the (3 + 1)-dimensional solutions of the appendix, up
to a conformal factor eΥ , since the latter metrics were Ricci flat by construction.
The embedding of the (1+1)-dimensional metric (5.5) into the conformally rescaled
(3 + 1)-dimensional solutions of the appendix are obtained by introducing the mass
parameter 2M (in units of G = c = 1) in the appropriate places in order to have
consistent units, and by writing
(dR/dr)2
   
2M r
eΥ(r) 1 − = tanh2 , eΥ(r) = 1, (5.8)
R(r) 2M 1 − 2M/R(r)
leading to the solutions for Υ(r) and R(r) respectively
 
1 r
eΥ = tanh2 , (5.9a)
1 − 2M/R(r) 2M
where
R − 2M
 
dR
Z
= R + 2M ln
1 − 2M/R 2M
 
dr r
Z
= = 2M ln sinh . (5.9b)
tanh r/2M 2M
This last equation (5.9b) yields the functional form R(r) (tortoise radial variable)
in implicit form for the radial function R(r). From Eq. (5.9b) one can infer that
R(r = 0) = 2M , R(r → ∞) → R ∼ r . (5.10)
1 The radial function R has a lower (ultraviolet cutoff) bound given by 2M . The fact
that the “point” r = 0 can have a nonzero proper area but zero volume seems to
3 indicate a “stringy” nature underlying the very notion of a point-mass itself. The
string worldsheet has area but no volume. Aspinwall27,28 has studied how a string
5 (an extended object) can probe space–time “points.”
Notice that if we allow for complex values of r, like r = 0 + i2M (π/2), that
furnish singularities in the metric (5.5), one must include a constant of integration
R0 = 2M (1 + iπ/2) to the solution in Eq. (5.9b):
  
R − 2M
  
iπ r
R − 2M 1 + + 2M ln = 2M ln sinh (5.11)
2 2M 2M
such that when one plugs in the value r = 0 + i2M (π/2) in the right-hand side of
7 Eq. (5.11), it coincides with the left-hand side of (5.11) when the value of the radial
function R(r = 0 + i2M π/2) = 2M (1 + iπ/2), after an analytical continuation
9 into the complex plane is performed. This is just a consequence of the relation
ln[sinh(iπ/2)] = ln[i sin(π/2)] = ln i = iπ/2.
11 This complex analytical continuation into regions where r, R are complex-valued
roughly speaking amounts to looking into the “interior” of the point-mass. Having
13 complex coordinates to probe into the “interior” of a point-mass is not so farfetched.
This suggests that quantum space–time might be intrinsically fractal , meaning that
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

On (2 + 2)-Dimensional Space–Times, Strings and Black Holes 17

1 the Hausdorff topological dimension of an object (let us say of a point) does not
coincide with the fractal dimension. For a throrough and profound treatment of
3 complex dimensions, fractal strings and the zeros of Riemman zeta function see
Ref. 100. The interplay among nonextensive statistics, chaos, complex dimensions,
5 logarithmic periodicity in the renormalization group and fractal strings see Ref. 101.
The conformal factor is
 
Υ 1 2 r
e = tanh , (5.12)
1 − 2M/R(r) 2M
where R(r) is given implicitly by (5.10). Notice that from the conditions in (5.10)
the conformal factor eΥ becomes unity at r = ∞ as it should if one wishes to
have asymptotic flatness. When r = 0 the conformal factor (5.12) is 00 undefined.
A careful study reveals that the conformal factor eΥ at r = 0 is zero so that
eΥ(r=0) R2 (r = 0) = 0 and the conformally rescaled proper area at r = 0 is zero.
Therefore, at r = 0 the conformally rescaled interval ds2 is zero consistent with the
fact that the (1+1)-dimensional metric exhibits a null horizon at r = 0. Concluding,
in this fashion, we have shown how one can embed the (1 + 1)-dimensional Bars–
Witten stringy black hole solution into the conformally rescaled (3 + 1)-dimensional
solutions of section of the appendix and are given by
 
2 2 r
ds = − tanh (dt)2 + (dr)2 + eΥ(r) R2 (r)dΩ2 . (5.13)
2M
Notice that the conformally rescaled metric (5.13) is not Ricci flat; it has singulari-
ties at complex values r = 0 + i2M π/2 ⇒ eΥ = ∞; R = 2M (1 + iπ/2) upon using
Eq. (5.11). There is a difference between the metric (5.13) with the Ricci flat metric
(outside the singularity at the point mass source) given in the Fronsdal–Kruskal–
Szekeres coordinates by
du dv
ds2 = −eW (u,v) + (R∗ (u, v))2 [sin2 φ(dθ)2 + (dφ)2 ]
1 − uv
du dv
= −eW (u,v) + (R∗ (u, v))2 dΩ2 , (5.14)
1 − uv
where W (u, v) and R∗ (u, v) are now two complicated functions of the two variables
7 u, v (since when one crosses the horizon the metric is no longer static). Whereas
in Eq. (5.13) one truly has a static metric everywhere and two functions of one
9 variable Υ(r), R(r) instead.
Before ending this work we will just add some remarks pertaining complex
11 gravity in 1 + 1 complex dimensions and its relation to ordinary gravity in 2 + 2
real dimensions. The properties of geometrical objects in the tangent space (at each
13 point of a curved space–time) associated to the complex, quaternionic and octo-
nionic algebra permits the construction of Einstein’s complexified, quaternionic and
15 octonionic gravity. In particular, gravity in 2 + 2 real dimensional can be studied
from the point of view of complex gravity in 1 + 1 complex dimensions. gravity in
17 4 + 4 real dimensional can be studied from the point of view of quaternionic gravity
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

18 C. Castro & J. A. Nieto

1 in 1 + 1 quaternionic dimensions, and gravity in 8 + 8 real dimensional can be seen

as octonionic gravity in 1 + 1 octonionic dimensions.102,103
To illustrate this, let us write the following complex line element in four complex-
dimensions:
dz1 dz1 + dz̃1 dz̃1 dz2 dz2 + dz̃2 dz̃2
ds2 = + . (5.15)
1 − z1 z1 − z̃1 z̃1 1 − z2 z2 − z̃2 z̃2
Complex gravity requires that gµν = g(µν) + ig[µν] so that now one has gνµ =
(gµν )∗ ,102–104 which implies that the diagonal components of the metric gz1 z1 =
gz2 z2 = gz̃1 z̃1 = gz̃2 z̃2 must be real, and which in turn implies that a real slice of
the 4-complex dimensional space spanned by the four complex variables z1 , z2 , z̃1 ,
z̃2 may be taken by imposing the following two constraints:
z̃1 = z1∗ , z̃2 = z2∗ (5.16)
and upon doing so one ends up with a four real -dimensional space of signature 2 + 2
whose real line element is
dz1 dz1 + dz1∗ dz1∗ dz2 dz2 + dz2∗ dz2∗
ds2 = + , (5.17)
1 − z 1 z1 − z 1 z1
∗ ∗ 1 − z2 z2 − z2∗ z2∗
where z1 , z2 are the complex coordinates of the 1 + 1 complex dimensional space–
time (2+2 real dimensional) while z1∗ , z2∗ are their complex conjugates, respectively.
After defining
1 1
z1 = √ (X + iT1 ) , z1∗ = √ (X − iT1 ) ,
2 2
(5.18)
1 1
z2 = √ (Y + iT2 ) , z2∗ = √ (Y − iT2 ) ,
2 2
3 the metric in Eq. (5.14) coincides precisely with the metric in Eq. (5.2) comprised of
the diagonal sum of two black hole solutions in 1 + 1 real dimensions. The quater-
5 nionic and octonionic versions of Eq. (5.16), in conjunction with the generalized
Einstein’s field equations, will be the subject of future investigations. The quater-
7 nionic analog of two-dimensional conformal field theory in four dimensions has been
studied by S. Vongehr.105 It is interesting to see (if possible) how one can construct
9 four-dimensional quantum nonlinear sigma models within the context of quantum 3-
branes (conformal field theories in the four-dimensional worldvolume of the 3-brane)
11 and find the analog of the coupled equations (5.7) associated with the vanishing of
the beta functions in two-dimensional CFT; namely from the perspective of a four-
13 dimensional quaternionic conformally invariant field theory formulated on Kulkarni
four-folds (the four-dimensional analog of Riemann surfaces) corresponding to 3-
15 branes moving in curved target space–time backgrounds. The cancellation of the
four-dimensional conformal anomaly should constrain the type of backgrounds on
17 which 3-branes can propagate.
It is worth mentioning that black hole solutions in a two times context have
19 been considered by some authors. In particular Kocinski and Wierzbicki107 con-
sidered Schwarzschild type solution in a Kaluza–Klein theory with two times. In
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

On (2 + 2)-Dimensional Space–Times, Strings and Black Holes 19

1 fact, using noncompactified Kaluza–Klein theory with internal signature of the

form 2 + 3 these authors determine a spherical symmetric solution. Vongehr108
3 also considered examples of black holes within the context of the two-times physics
formulation of Bars (see Ref. 106 and references therein). Their basic examples
5 coreponds essentially to a solutions associated with the signatures 1 + 1 and 2 + 3.
Finally, the four-dimensional Kaluza–Klein approach to general relativity in
7 2 + 2 as a local product of a (1 + 1)-dimensional base manifold and a (1 + 1)-
dimensional fiber space109,110 warrants further investigation in so far that 2 + 2
9 gravity can be described by a (1 + 1)-dimensional Yang–Mills gauge theory of dif-
feormorphims of the two-dimensional fiber space coupled to a (1 + 1)-dimensional
11 nonlinear sigma model and a scalar field; i.e. this formulation of 2 + 2 gravity
by109,110 is more closely related to the stringy picture of the Bars–Witten black
13 hole in 1 + 1-dimensions. Thus, it seems interesting to pursue further research to
see the possible connection between the present work and these other approaches.
15 For example, to study black holes solutions in noncommutative geometry,73 in par-
ticular Finsler spaces,67–71,79,80 phase spaces74–78,81,82 and the implications of the
17 minimal Planck scale41 stringy uncertainty relations83,84 in black holes physics.86–89

Appendix A. Schwarzschild-like Solutions in Any Dimension D > 3

Let us start with the line element

ds2 = −eµ(r) (dt1 )2 + eν(r) (dr)2 + R2 (r)g̃ij dξ i dξ . (A.1)

Here, the metric g̃ij corresponds to a homogeneous space and i, j = 3, 4, . . . , D − 2.

The only nonvanishing Christoffel symbols are
1 0 1 0 1 0 µ−ν
Γ121 = µ , Γ222 = ν , Γ211 = µe ,
2 2 2
(A.2)
R0 i
Γ2ij = −e −ν 0
RR g̃ij , Γi2j = δ , Γijk = Γ̃ijk ,
R j
and the only nonvanishing Riemann tensor are
1 1 1
R1212 = − µ00 − µ02 + ν 0 µ0 ,
2 4 4
1
R1i1j = − µ0 e−ν RR0 g̃ij ,
2
 
2 µ−ν 1 00 1 02 1 0 0 (A.3)
R121 = e µ + µ − νµ ,
2 4 4
 
1 0
R2i2j = e−ν ν RR0 − RR00 g̃ij ,
2
Rijkl = R̃jkl
i
− R02 e−ν δki g̃jl − δli g̃jk .


March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

20 C. Castro & J. A. Nieto

The field equations are


1 02 1 0 0 (D − 2) 0 R0

µ−ν 1 00
R11 = e µ + µ − µν + µ = 0, (A.4)
2 4 4 2 R
 
1 00 1 02 1 0 0 1 0 R0 R00
R22 = − µ − µ + µ ν + (D − 2) ν − = 0, (A.5)
2 4 4 2 R R
and
 
e−ν 1 0
Rij = (ν − µ0 )RR0 − RR00 − (D − 3)R02 g̃ij
R2 2
k
(D − 3)g̃ij = 0 ,
+ (A.6)
R2
where k = ±1, depending if g̃ij refers to positive or negative curvature. From the
combination e−µ+ν R11 + R22 = 0 we get
2R00
µ0 + ν 0 = . (A.7)
R0
The solution of this equation is
µ + ν = ln R02 + a , (A.8)
1 where a is a constant.
Substituting (A.7) into Eq. (A.6) we find
e−ν (ν 0 RR0 − 2RR00 − (D − 3)R02 = −k(D − 3) (A.9)
or
γ 0 RR0 + 2γRR00 + (D − 3)γR02 = k(D − 3) , (A.10)
where
γ = e−ν . (A.11)
The solution of (A.10) for an ordinary D-dimensional space–time (one temporal
dimension) corresponding to a (D − 2)-dimensional sphere for the homogeneous
space can be written as
  −2
16πGD M dR
γ = 1−
(D − 2)ΩD−2 RD−3 dr
 −1  2
ν 16πGD M dR
⇒ grr = e = 1 − , (A.12)
(D − 2)ΩD−2 RD−3 dr
where ΩD−2 is the appropriate solid angle in (D − 2)-dimensional and GD is the
3 D-dimensional gravitational constant whose units are (length)D−2 . Thus GD M
has units of (length)D−3 as it should. When D = 4 as a result that the two-
5 dimensional solid angle is Ω2 = 4π one recovers from Eq. (A.12) the four-
dimensional Schwarzchild solution. The solution in Eq. (A.12) is consistent with
 1st Reading
March 9, 2007 10:44 WSPC/139-IJMPA 03619

On (2 + 2)-Dimensional Space–Times, Strings and Black Holes 21

1 Gauss law and Poisson’s equation in D − 1 spatial dimensions obtained in the

Newtonian limit.
For the most general case of the (D − 2)-dimensional homogeneous space we
should write
βD G D M
−ν = ln(k − ) − 2 ln R0 , (A.13)
RD−3
where βD is a constant. Thus, according to (A.8) we get
 
βD G D M
µ = ln k − + const (A.14)
RD−3
we can set the constant to zero, and this means the line element (A.1) can be
written as
 
2 βD G D M
ds = − k − (dt1 )2
RD−3
(dR/dr)2
+  (dr)2 + R2 (r)g̃ij dξ i dξ . (A.15)
βD GD M
k − RD−3

3 One can verify, taking for instance (A.5), that Eqs. (A.4)–(A.6) do not determine
the form R(r). It is also interesting to observe that the only effect of the homo-
5 geneous metric g̃ij is reflected in the k = ±1 parameter, associated with a positive
(negative) constant scalar curvature of the homogeneous (D −2)-dimensional space.

7 Acknowledgments
We wish to thank the referee for his numerous and critical suggestions to improve
9 this work. J. A. Nieto thanks L. Ruiz, J. Silvas and C. M. Yee for helpful comments.
This work was partially supported by grants PIFI 3.2. C. Castro thanks M. Bowers
11 for hospitality and Sergiu Vacaru for many discussions about Finsler geometry and
related topics.

13 References
1. M. F. Atiyah and R. S. Ward, Commun. Math. Phys. 55, 117 (1977).
15 2. R. S. Ward, Phil. Trans. R. Soc. London, Ser. A 315, 451 (1985).
3. M. A. De Andrade, O. M. Del Cima and L. P. Colatto, Phys. Lett. B 370, 59 (1996),
17 hep-th/9506146.
4. M. A. De Andrade and O. M. Del Cima, Int. J. Mod. Phys. A 11, 1367 (1996).
19 5. M. Carvalho and M. W. de Oliveira, Phys. Rev. D 55, 7574 (1997), hep-th/9612074.
6. H. Garcia-Compean, N = 2 string geometry and the heavenly equations, in Proc.
21 Conf. on Topics in Mathematical Physics, General Relativity and Cosmology in
Honor of Jerzy Plebanski, Mexico City, Mexico, 17–20 Sept. 2002, eds. H. Garcı́a-
23 Compeán et al. (World Scientific, 2006), hep-th/0405197.
7. H. Ooguri and C. Vafa, Nucl. Phys. B 367, 83 (1991).
25 8. H. Ooguri and C. Vafa, Nucl. Phys. B 361, 469 (1991).
 1st Reading
March 9, 2007 10:44 WSPC/139-IJMPA 03619

22 C. Castro & J. A. Nieto

1 9. D. Kutasov, E. J. Martinec and M. O’Loughlin, Nucl. Phys. B 477, 675 (1996),

hep-th/9603116.
3 10. J. Martinec, Matrix theory and N = (2, 1) strings, hep-th/9706194.
11. D. Kutasov and E. J. Martinec, Nucl. Phys. B 477, 652 (1996), hep-th/9602049.
5 12. D. Kutasov and E. J. Martinec, Class. Quantum Grav. 14, 2483 (1997), hep-
th/9612102.
7 13. P. G. O. Freund, Introduction to Supersymmetry (Cambridge University Press, 1986).
14. S. V. Ketov, H. Nishino and S. J. Gates, Phys. Lett. B 307, 323 (1993), hep-
9 th/9203081.
15. J. A. Nieto, M. P. Ryan, O. Velarde and C. M. Yee, Int. J. Mod. Phys. A 19, 2131
11 (2004), hep-th/0401145.
16. J. A. Nieto, Nuovo Cimento B 120, 135 (2005), hep-th/0410003.
13 17. M. B. Green, Nucl. Phys. B 293, 593 (1987).
18. P. K. Townsend, Phys. Lett. B 373, 68 (1996), hep-th/9512062.
15 19. J. A. Nieto, Mod. Phys. Lett. A 10, 3087 (1995), gr-qc/9508006.
20. J. A. Nieto, From the Quatl to the Ketzal of a generalized string theory, prepared
17 for 5th Mexican Workshop on Particles and Fields, Puebla, Mexico, 30 October–3
November 1995, published in Particles and Fields and Phenomenology of Fundamen-
19 tal Interactions, Puebla, 1995, 525-528.M.
21. J. Duff, Phys. Lett. B 641, 335 (2006), hep-th/0602160.
21 22. J. A. Nieto, The 2 + 2 signature and the 1 + 1 matrix-brane, hep-th/0606219.
23. C. Castro, J. Math. Phys. 35, 920 (1994).
23 24. C. Castro, J. Math. Phys. 35, 3013 (1994).
25. C. Castro, Phys. Lett. B 288, 291 (1992).
25 26. C. Castro, J. Math. Phys. 34, 681 (1993).
27. P. S. Aspinwall, J. High Enegy Phys. 0407, 021 (2004), hep-th/0312188.
27 28. P. S. Aspinwall, J. High Energy Phys. 0301, 002 (2003), hep-th/0203111.
29. L. Abrams, Can. J. Phys. 67, 919 (1989).
29 30. L. Abrams, Phys. Rev. D 20, 2474 (1979).
31. L. Abrams, Phys. Rev. D 21, 2438 (1980).
31 32. L. Abrams, Phys. Rev. D 21, 2941 (1980).
33. C. Castro, Phys. Lett. B 626, 209 (2005).
33 34. C. Castro, Found. Phys. 35, 971 (2005).
35. C. Castro, Progr. Phys. 2, 86 (2006).
35 36. C. Castro, Exact solutions of Einstein’s equations in the presence of delta function
point mass source, submitted to Found. Phys.
37 37. C. Castro, Mod. Phys. Lett. A 17, 2095 (2002).
38. C. Castro, On dark energy, Weyl geometry and different derivations of the vacuum
39 energy density, to appear in Found. Phys., 2007.
39. C. Castro, On novel static spherically symmetric solutions of Einstein equations and
41 the cosmological constant problem, CTSPS preprint, May 2006.
40. T. Padmanabhan, Dark energy: Mystery of the millennium, astro-ph/0603114.
43 41. L. Nottale, Fractal Spacetime and Microphysics: Towards Scale Relativity (World
Scientific, Singapore, 1992).
45 42. P. Fiziev, Gravitational field of massive point particle in general relativity, gr-
qc/0306088.
47 43. P. Fiziev and S. V. Dimitrov, Point electric charge in general relativity, hep-
th/0406077.
49 44. P. Fiziev, The gravitational field of massive non-charged point source in general
relativity, gr-qc/0412131.
 1st Reading
March 9, 2007 10:44 WSPC/139-IJMPA 03619

On (2 + 2)-Dimensional Space–Times, Strings and Black Holes 23

1 45. P. Fiziev, On the solutions of Einstein equations with massive point source, gr-
qc/0407088.
3 46. A. Einstein, Sitzungsber. Preuss. Akad. Berlin II, 831 (1915).
47. K. Schwarzschild, Sitzungsber. Preuss. Akad. Berlin I, 189 (1916) [English transla-
5 tions: S. Antoci and A. Loinger, physics/9905030].
48. M. Brillouin, J. Phys. Rad. 23, 43 (1923) [English translation: S. Antoci,
7 physics/0002009].
49. D. Hilbert, Nachr. Ges. Wiss Gottingen Math. Phys. K 1, 53 (1917).
9 50. H. Weyl, Ann. Physik (Leipzig) 54, 117 (1917).
51. J. Droste, Proc. Ned. Akad. West, Ser. A 19, 197 (1917).
11 52. A. Loinger, On Black Holes and Gravitational Waves (La Goliardica Pavese, 2002),
129 pages.
13 53. A. Loinger and T. Marsico, On the gravitational collapse of a massive star,
physics/0512232.
15 54. S. Crothers, Progr. Phys. 1, 68 (2005).
55. S. Crothers, Progr. Phys. 2, 3 (2005).
17 56. S. Crothers, Progr. Phys. 3, 7 (2005).
57. N. Stavroulakis, Progr. Phys. 2, 68 (2006).
19 58. M. Pavsic, Obzornik za Matematiko in Fiziko 28, 5 (1981) [English translation: Grav-
itational field of a point source].
21 59. M. Ibison, private communication.
60. C. Fronsdal, Phys. Rev. 116, 778 (1959).
23 61. M. Kruskal, Phys. Rev. 119, 1743 (1960).
62. G. Szekers, Publ. Mat. Debreca 7, 285 (1960).
25 63. J. F. Colombeau, New Generalized Functions and Multiplcation of Distributions
(North Holland, Amsterdam, 1984).
27 64. J. F. Colombeau, Elementary Introduction to Generalized Functions (North Holland,
Amsterdam, 1985).
29 65. J. Heinzke and R. Steinbauer, Remarks on the distributional Schwarzschild geometry,
gr-qc/0112047.
31 66. M. Grosser, M. Kunzinger, M. Oberguggenberger and R. Steinbauer, Geometric
Theory of Generalized Functions with Applications to Relativity, Kluwer Series on
33 Mathematics and its Applications, Vol. 537 (Kluwer, Dordrecht, 2001).
67. S. Vacaru, P. Stavrinos, E. Gaburov and D. Gonta, Clifford and Riemann–Finsler
35 structures in geometric mechanics and gravity, to appear in Book Title (Geometry
Balkan Press), 693 pages.
37 68. S. Vacaru, Phys. Lett. B 498, 74 (2001).
69. S. Vacaru, J. Math. Phys. 46, 042503 (2005).
39 70. S. Vacaru, J. Math. Phys. 46, 032901 (2005).
71. S. Vacaru, J. Math. Phys. 47, 093504 (2006).
41 72. S. Antoci and D. E. Liebscher, Reinstating Schwarzschild’s original manifold and its
singularity, gr-qc/0406090.
43 73. J. Wess, Einstein–Riemann gravity on deformed spaces, hep-th/0611025.
74. M. Born, Proc. R. Soc. A 165, 291 (1938).
45 75. E. Caianiello, Lett. Nuovo Cimento 32, 65 (1981).
76. G. Lambiase, G. Papini and G. Scarpetta, Phys. Lett. A 263, 147 (1999).
47 77. M. Toller, Int. J. Theor. Phys. 29, 963 (1990).
78. V. Nesterenko, Class. Quantum Grav. 9, 1101 (1992).
49 79. H. Brandt, Contemp. Math. 196, 273 (1996).
80. H. Brandt, Chaos, Solitons & Fractals 10, 267 (199???).
March 9, 2007 10:44 WSPC/139-IJMPA 03619
1st Reading

24 C. Castro & J. A. Nieto

1 81. C. Castro, Found. Phys. 35, 971 (2005).

82. C. Castro, Progr. Phys. 1, 20 (2006).
3 83. D. Amati, M. Ciafaloni and G. Veneziano, Phys. Lett. B 197, 81 (1987).
84. D. Gross and P. Mende, Phys. Lett. B 197, 129 (1987).
5 85. C. Castro, J. Phys. A: Math. Gen. 39, 14205 (2006).
86. K. Nozari and S. Mehdipour, Failure of standard thermodynamics in Planck scale
7 black hole system, hep-th/061007.
87. L. Colatto, A. Penza and W. Santos, Noncommutative geometry induced by spin
9 effects, hep-th/0512266.
88. F. Nasseri and S. Alavi, Schwarzschild black hole in noncommuttaive spacetime,
11 hep-th/0508051.
89. A. Lewis, Coulomb potential of a point mass in theta noncommutative geometry,
13 hep-th/0605140.
90. A. Bonanno and M. Reuter, Renormalization group improved black hole spacetime,
15 hep-th/0002196.
91. M. Reuter and J. M. Schwindt, A minimal length from cutoff modes in asymptotically
17 safe quantum gravity, hep-th/0511021.
92. M. Reuter and J. M. Schwindt, Scale-dependent structures and causal structures in
19 quantum Einstein gravity, hep-th/0611294.
93. A. Bonanno and M. Reuter, Phys. Rev. D 73, 0830005 (2006), hep-th/0602159.
21 94. C. Castro, A. Nieto and J. F. Gonzalez, Running Newtonian coupling and horizonless
solutions in Quantum Einstein Gravity, to appear in Quantization in Astrophysics,
23 Brownian Motion and Supersymmetry, eds. F. Smarandache and V. Christianato
(MathTiger Publishers, Chennai, India, 2006).
25 95. I. Bars, Lecture at Strings 91, Stonybrook, June 1991.
96. E. Witten, Phys. Rev. D 44, 314 (1991).
27 97. E. Witten, On black holes in string theory, Lecture given at Strings 1991 Conf.,
Stony Brook, New York, Jun 1991, hep-th/9111052.
29 98. S. Ketov, Quantum Non-Linear Sigma Models, Conformal Field Theory, Supersym-
metry, Black Holes and Strings (Springer Verlag, Berlin-Heidelberg, 2000).
31 99. M. Green, J. Schwarz and E. Witten, Superstring Theory (Cambridge University
Press, 1986).
33 100. M. Lapidus and M. Frankenhuysen, Complex Dimensions and the Zeros of the Zeta
Functions (Birkhauser, New York, 2000).
35 101. C. Castro, Physica A 347, 184 (2005).
102. S. Marques and C. Oliveira, Phys. Rev. D 36, 1716 (1987).
37 103. K. Borchsenius, Phys. Rev. D 13, 2707 (1976).
104. J. Moffat and D. Boal, Phys. Rev. D 11, 1375 (1975).
39 105. R. Zucchini, J. Geom. Phys. 27, 113 (1998).
106. I. Bars, The standard model of particles and forces in the framework of 2T-physics,
41 hep-th/0606045.
107. J. Kocinski and M. Wierzbicki, Rel. Grav. Cosmol. 1, 19 (2004), gr-qc/0110075.
43 108. S. Vongehr, Examples of black holes in two time physics, hep-th/9907077.
109. J. H. Yoon, 4-dimensional Kaluza–Klein approach to general relativty in 2 + 2 space-
45 times, gr-qc/9611050.
110. J. H. Yoon, Algebraically special class of spacetimes and 1 + 1-dimensional field
47 theories, hep-th/9211129.